Increased expression of Drosophila Sir 2 extends life span in a dose-dependent manner

dSir2 expression from UAS-dSir2 transgenes can increase or decrease life span

To examine the effect of increasing dSir2 expression on longevity, we examined the life span of flies carrying either of two newly generated UAS-dSir2 transgenes (Sir2-1 and Sir2-3) driven by the ubiquitous Tubulin-Gal4 driver. The life span of these flies was compared to Tubulin-Gal4 flies crossed to wild type flies from the same genetic background as the UAS-dSir2 transgenes. Under these conditions, Tubulin-Gal4>Sir2-3 flies showed life span extension, while Tubulin-Gal4>Sir2-1 flies had a shortened life span (Fig. 1A). The discrepancy of finding of a decrease in life span with the Sir2-1 transgene and an increase with the Sir2-3 transgene led us to measure the levels of dSir2 expression in these flies. Due to difficulties in establishing a functional dSIR2 western protocol, we instead utilized RT-qPCR to measure mRNA levels in this study. We found that flies containing the Sir2-1 transgene over expressed dSir2 at very high levels (~ 45-fold increase), compared to the Sir2-3 transgene (~ 11-fold increase) (Fig. 1A, inset).

To eliminate the possibility that increased expression of dSir2 during development in the presence of the constitutively expressed Tubulin-Gal4 driver might contribute to the deleterious effect we see on adult life span in the Sir2-1 line, we utilized the RU-486 inducible Gene Switch system in which control and experimental flies are genetically identical offspring from the same cohort and increased expression can be induced in, and thus restricted only to adult life. We examined the life span of flies with increased expression of dSir2 from the Sir2-1 transgene driven by the conditional Tubulin Gene Switch driver (Tubulin-Gene Switch). When expression of the Sir2-1 construct was induced using RU-486 in the food of adult flies, the life span of these flies was shortened, as seen for the Tubulin-Gal4>Sir2-1 life span (Fig. 1B). We measured the level of dSir2 mRNA in these flies, and found an approximately 50-fold induction of dSir2, consistent with the high level of induction seen with the constitutive Tubulin-Gal4 driver (Fig. 1B, inset).

Additionally, we crossed another UAS-dSir2 line, Sir2-4, to the Tubulin Gene Switch driver and observed an increase of ~15-fold in dSir2 levels for RU-486-induced flies compared to EtOH controls, without any observable change in median life span (Fig. S1).

Given that the level of dSir2 induction using both the Sir2-1 and Sir2-4 transgenes in these studies was high, we hypothesized that these levels of dSir2 might be detrimental to life span. In order to better test the effects of increased dSir2 expression at more moderate levels, we performed life span and mRNA expression experiments using the Sir2-3 transgene, which induced dSir2 expression more moderately than Sir2-1. We also included an additional genetic background control in this experiment, where we crossed w¹¹¹⁸ flies (injection stock for the transgene lines Sir2-1, Sir2-3, Sir2-4) to the Tubulin-Gene Switch driver, and measured life span in these flies with and without RU-486 induction. Life spans of control w¹¹¹⁸ flies carrying the Tubulin-Gene Switch driver was the same with or without RU-486, demonstrating that RU-486 treatment does not directly affect life span in this w¹¹¹⁸ strain.

When the Sir2-3 transgene was expressed ubiquitously using the Tubulin-Gene Switch driver in the presence of RU-486, we observed no life span extension compared with genetically identical uninduced (ethanol-fed) flies. However, inclusion of the additional w¹¹¹⁸ genetic background control in this study allowed for a comparison of uninduced control flies (Tubulin-Gene Switch> Sir2-3) with the w¹¹¹⁸ controls. Surprisingly, we found that in comparison to the Tubulin-Gene Switch>w¹¹¹⁸ controls, both the uninduced and the induced Tubulin-Gene Switch>Sir2-3 flies exhibited an extended life span (Fig. 1C). This result was unexpected, given that the absence of RU-486 in the food should have meant that there was no induction of dSir2. In order to clarify this result, we also examined the levels of dSir2 mRNA in all of these flies, and found that the diluent, ethanol-fed control Tubulin-Gene Switch>Sir2-3 flies had a mild but significant elevation in dSir2 levels (~1.5-fold induction) when compared with the Tubulin-Gene Switch>w¹¹¹⁸ controls. The Tubulin-Gene Switch>Sir2-3 flies that were induced with RU-486 in the food showed an ~ 11-fold increase in dSir2 levels (Fig. 1C, inset). This finding suggests that in these conditions, a modest ~1.5-fold and significantly higher ~11-fold elevation of dSir2 can extend life span.

Finally, we crossed the Sir2-3 transgenic flies to the neuron-specific ELAV-Gene Switch driver and evaluated life span extension and dSir2 mRNA levels. Again, we included the ELAV-Gene Switch>w¹¹¹⁸ genetic background control to allow for comparison with the uninduced ELAV-Gene Switch>Sir2-3 ethanol controls. In this experiment, the uninduced ELAV-Gene Switch>Sir2-3 flies did not have significantly elevated dSir2 levels compared to the ELAV-Gene Switch>w¹¹¹⁸ genetic control. Consistent with this finding, we saw no life span extension for the uninduced ELAV-Gene Switch>Sir2-3. However, the RU-486 induced ELAV-Gene Switch>Sir2-3 flies had a 5-fold increase in dSir2 mRNA levels and extended life span as compared to the uninduced Elav-Gene Switch>Sir2-3 control flies and the Elav-Gene Switch>w¹¹¹⁸ control flies (Fig. 1D).

Moderate increases in dSir2 expression from UAS-dSir2 transgenes extend life span

The ability of the Sir2-3 line to induce dSir2 expression in the uninduced condition of the Tubulin-Gene Switch life span led us to examine whether the additional copy of dSir2 from the UAS-dSir2 transgene may under certain circumstances express dSir2 in the absence of Gal4. To directly test whether the presence of one copy of the UAS-dSir2 transgene could lead to increased expression of dSir2 without the presence of a Gal4 driver, we crossed three different homozygous UAS-dSir2 transgenic lines (Sir2-1, Sir2-3, and Sir2-4) to their genetic background control (w¹¹¹⁸) and measureddSir2 mRNA levels in the F1 heterozygotes (Fig. 1E, inset). One copy of the Sir2-1 insertion (Sir2-1 / +) did not result in significantly elevated dSir2 levels, but both Sir2-3 / + and Sir2-4 / + exhibited significantly increased dSir2 mRNA levels (~ 2-fold increase over w¹¹¹⁸dSir2 levels). This result suggests that the life span extension observed in the control condition of the Tubulin-Gene Switch>Sir2-3 and Sir2-4 life spans may be due to uninduced expression of the dSir2 construct under some conditions.

In flies from the same cohort as those used for qPCR, we tested whether dSir2 expression from the UAS-dSir2 transgenes in the absence of a Gal4 driver was sufficient to extend life span (Fig. 1E). A single copy of Sir2-1 that did not show elevated dSir2 mRNA levels did not extend lifespan, whereas one copy of either Sir2-3 or Sir2-4 did increase lifespan, compared to the genetically matched w¹¹¹⁸ controls. These results suggest that moderately increased dSir2 expression, here from un-driven UAS-dSir2 constructs, can extend lifespan in flies.

dSir2-mediated life span extension is dose-dependent

The extension of life span seen when dSir2 expression is moderately increased led to the hypothesis that there may be a dose-dependent element to the effect of dSir2 expression on life span. Previous studies showing that increased dSir2 expression extends life span used fly lines that expressed dSir2 from its endogenous locus at relatively modest levels (~3-5 fold increase) [5,6]. We noted that the flies with increased dSir2 expression from the Sir2-3 and Sir2-4 transgenes express dSir2 to a similar extent, from 2-5 fold increased, when life span is extended, while some of the conditions that we had tested had much higher levels of dSir2, and rarely extended life span. We noted that the recent study demonstrating no effect of increasing dSir2 on life span reported that their induced dSir2 levels were either 1.5- fold or 8-fold increased [10]. In order to examine whether increased dSir2 expression might have a dose-dependent effect on life span, we plotted dSir2 expression levels and life span extension from both our studies and those of published life spans (Fig. 2A and B).

As can be seen in Fig. 2, increased dSir2 expression is associated with life span extension when mRNA levels are elevated approximately 2 to 5-fold compared to controls. In addition, very high dSir2 expression (greater than 30-fold increase) is associated with shortened life span, suggesting that such high levels of dSir2 may have a toxic effect. Interestingly, increased expression of dSir2 in the dSir2^EP2300 line extends life span more reliably and to a greater extent than induction at similar levels from the UAS-dSir2 transgenes. This suggests that the presence of endogenous regulatory elements may be beneficial. Given that all of the dSir2 expression levels for flies studied in [10] were outside the predicted optimal range that we report here, we hypothesized that we should be able to extend life span with these transgenic flies if dSir2 expression within our predicted optimal window could be achieved.

To test this hypothesis we used the Gene Switch system to increase expression of dSir2 from the myc-tagged UAS-dSir2 transgenic fly lines used in [10]. The Tubulin-Gene Switch driver permitted us to express dSir2 in the same tissues as in the original study, and to limit the expression to only adult life, avoiding any potential deleterious effects of increased dSir2 during development. At the same time, this system allowed us to vary the concentration of RU-486 in the food, thereby titrating dSir2 expression levels in these flies. We used three concentrations of RU-486 (100, 200, and 500 μM) to induce dSir2 expression, and measured the dSir2 transcript levels for each of these conditions. We then plotted the median life span increase for both our own new life spans and those previously published with these constructs as a function of the measured dSir2 levels (Fig. 2C). Life span was extended to the greatest extent when dSir2 was expressed at moderate levels. These results demonstrate a dose-dependent response for increased dSir2 expression on fly life span, even for transgenic constructs previously reported not to extend life span.

High levels of dSir2 expressed from transgenes induce cellular toxicity and dnaJ-H expression

The UAS insertion in the dSir2^EP2300 over-expression line is located within the 5' UTR of dSir2, oriented in a direction so as to specifically drive expression of dSir2 and not the adjacent gene dnaJ-H [9]. Although one study concluded that increased dSir2 transgene expression is lethal and that dnaJ-H suppresses this phenotype [12], in a subsequent study using the same dSir2^EP2300 line driven by ELAV-Gene Switch at two different doses of RU-486, a dose-dependent increase in life span was observed coincident with increased dSir2 but not dnaJ-H levels [6]. Similarly, a life span extending increase in dSir2 expression from dSir2^EP2300 in fat body using the S₁-106 Gene Switch driver showed no increase in dnaJ-H expression [7]. These conflicting studies led us to test whether the expression of dnaJ-H in the eye might be unrelated to expression of dSir2 by the dSir2^EP2300 EP element.

We measured dnaJ-H and dSir2 transcript levels in the heads of ELAV-Gene Switch>dSir2^EP2300 flies and observed that induction of dSir2^EP2300 from the native dSir2 locus as in [5] leads to a 3.5-fold increase in dSir2 and no increase in dnaJ-H levels (Fig. S2). We also measured dSir2 and dnaJ-H transcripts in three of our own UAS-dSir2 transgene containing fly lines (Sir2-1, Sir2-3, and Sir2-4), driven by the ELAV-Gal4 driver (Fig. 3A). Expression of dnaJ-H is increased in some of the UAS-dSir2 transgenic induced expression conditions but only when the level of dSir2 induction is greater than 6-fold (Fig. 3A, diamonds).

In order to explore the mechanism by which dnaJ-H is elevated in response to transgenically expressed dSir2, we measured the level of expression of the JNK signaling gene, puc phosphatase (puc), in these conditions and found that puc follows the same pattern as dnaJ-H, exhibiting increased expression when dSir2 reaches high levels (Fig. 3A, squares). Thus, expression of dSir2 at high levels may induce dnaJ-H as part of a cellular stress response, not because it is located adjacent to dSir2, as was suggested in a previous report [12]. Importantly, moderate levels of increased dSir2, such as those typically seen under conditions associated with life span extension, do not induce dnaJ-H expression.

These results led us to consider that the lethality previously observed in [12] might have resulted from toxically high levels of dSir2. Since crossing the UAS-dSir2 line from that study (referred to here as Sir2(G)) [12] to constitutive GAL4 drivers results in developmental and cellular lethality, we used the ELAV-Gene Switch driver to induce expression from the Sir2(G) transgene in neurons for 10 days during adult life, and measured dSir2 levels in fly heads. We compared dSir2 expression from the ELAV-Gene Switch>Sir2(G) flies with ELAV-Gene Switch>Sir2-1 and ELAV-Gene Switch>Sir2-3 flies under the same conditions, and found that Sir2-3, which has moderate levels of increased expression, and Sir2-1, which shows very high dSir2 levels, both express dSir2 at much lower levels than the Sir2(G) line, which showed a 50-fold increase in mRNA levels under these conditions (Fig. 3B). We conclude that the lethality of the Sir2(G) transgene resulted from extremely high dSir2 levels.

Since the Sir2(G) transgene expresses dSir2 at very high levels, and dnaJ-H is elevated in our highly expressing dSir2 transgenes, we would predict that under conditions of high dSir2 induction, the Sir2(G) transgene should also induce dnaJ-H. Additionally, we expect that induction of dSir2 expression from Sir2(G) at lower levels should not induce dnaJ-H. To test this, we crossed the lethal Sir2(G) line to the ubiquitous Tubulin-Gene Switch driver and induced dSir2 at different levels by using two different doses of RU-486 (Fig. 3C). At a low concentration of RU-486 (50 μM), dSir2 is increased 3-fold and dnaJ-H levels are not elevated, while at a higher concentration of RU-486 (200 μM), dSir2 is induced 25-fold over control levels and dnaJ-H expression is increased almost 3-fold. Therefore, the Sir2(G) transgene induces dnaJ-H induction when dSir2 levels are high, but not when dSir2 is slightly induced.

Finally, we tested whether dnaJ-H induction was specific to elevated dSir2 by expressing the foreign GFP mRNA using the strong, ubiquitous, constitutive daughterless-Gal4 driver and measuring dnaJ-H levels (Fig. S3). Interestingly, expressing GFP with this driver led to a 2.5-fold increase in dnaJ-H. The induction of a strong dnaJ-H expression response when proteins other than dSir2 are expressed indicates that the induction of dnaJ-H when dSir2 is highly over-expressed is likely part of the cellular stress response itself, and not due to a specific function of dSir2.

Drosophila Stocks and Maintenance

All flies were maintained at 25°C in a temperature-controlled incubator at 50% humidity with a 12-hour light/dark cycle. The Tubulin-Gal4 (5138), ELAV-Gal4 (Bloomington 458), and Daughterless-Gal4 (Bloomington 5460) lines were from Bloomington Stock Center. ELAV Gene Switch was a gift from H. Keshishian (Yale University, New Haven, CT). dSir2^EP2300 from the Bloomington stock center was backcrossed 9 times into the w¹¹¹⁸ control line before use in qPCR studies. The lethal UAS-dSir2(G) line was a kind gift from KT Min (Indiana University, Bloomington, IN). The Tubulin Gene Switch and Myc-tagged UAS-dSir2 lines (Myc2 and Myc9) were kind gifts from S. Pletcher (University of Michigan, Ann Arbor, MI).

Construction of UAS-dSir2 lines

To generate the UAS-dSir2 lines (Sir2-1, Sir2-4 and Sir2-4), full length dSir2 was cloned from an adult Drosophila cDNA library into pDONR221 (Invitrogen) and then the pTW vector (Drosophila Gateway Vector Collection, Carnegie Institute) using the Gateway system. Primers used were dSir2-GW-F: GGGGACAAGTTTGTACAAAAAAGC AGGC TGCACCATGATGGAAAATTACGAGGAA, and dSir2-GW-R: GGGGACCACTTTGTACAAGAA AGCTGGGTCCACTGCTGCTAACTGTCCTGG, amplifying a 2.4 kb fragment. Embryo injection of pTW-dSir2 into the w¹¹¹⁸ background was performed by Bestgene, Inc. (Chino Hills, CA). Flies were selected for germline transformation, and kept as homozygous stocks. Different lines were established from three separate integration events.

Life span Assays

All life span experiments were performed on food containing 150 g/L sucrose, 150 g/L autolysed yeast, 20 g/L cornmeal, and 20 g/L agar, all w/v. Life span experiments using UAS-Sir2 Myc2 or UAS-Sir2 Myc9 lines were performed on food containing 150 g/L dextrose, 150 g/L autolysed yeast, 20 g/L cornmeal, and 20 g/L agar, all w/v. Life span experiments using Gene Switch drivers were conducted on food containing these ingredients plus either RU-486 (Cayman Chemical, Ann Arbor, MI) in ethanol at the stated concentration, or the diluent alone (20 ml/L ethanol) for the control. Flies were collected under light CO₂ anesthesia, randomly divided into treatment groups, and housed at a density of 25 males and 25 females per vial. Flies were passed onto fresh food every other day (Experiments labeled “EOD” in Table S1) or every day (labeled “ED” in Table S1), and the number and sex of dead flies recorded.

RNA Isolation and qPCR

Flies for qPCR analysis were collected in a 24-hour window and grown under the same conditions as flies in life span experiments. They were passed onto new food every other day (or every day for life spans passed every day), and on day 10, were snap-frozen in liquid nitrogen and stored at −80°C. For mRNA isolation, 30 whole male flies (for ubiquitious driver experiments) or male fly heads (for ELAV-Gal4 and ELAV-Gene Switch experiments) were used to isolate mRNA using the Dynabeads mRNA DIRECT kit (Life Technologies). mRNA was reverse transcribed using the iScript cDNA Synthesis Kit (Bio-Rad). 50 ng of cDNA was used as qPCR template. qPCR was performed on a ABI 7500 FAST Real-Time PCR machine, using SYBR Green PCR Master Mix (Life Technologies). Each qPCR run was performed using three biological replicates in triplicate. Primer sequences used for amplification can be found in Table S4. GAPDH was used as a normalizing gene.

Statistical Analysis

All life spans were analyzed in a Log Rank test using the Survival package in R. Student's t-test was used to analyze the statistical significance of qPCR data.